The microenvironment of growth factors in the subependymal zone (SEZ) of the adult brain provides the instructive milieu for neurogenesis to proceed in this germinal niche. In particular, tight regulation of bone morphogenetic protein (BMP) signaling is essential to balance proliferative and non-proliferative cell fate specification. However, the regulatory pathways that control BMP signaling in the SEZ are still poorly defined. We demonstrate that LRP2, a clearance receptor for BMP4 is specifically expressed in ependymal cells of the lateral ventricles in the adult brain. Intriguingly, expression is restricted to the ependyma that faces the stem cell niche. Expression is not seen in ependyma elsewhere in the lateral ventricles or in the dentate gyrus, the second major neurogenic zone of the adult brain. We further show that lack of LRP2 expression in adult mice results in impaired proliferation of neural precursor cells in the SEZ resulting in decreased numbers of neuroblasts reaching the olfactory bulb. Reduced neurogenesis coincides with increased BMP4 expression and enhanced activation of downstream mediators phospho-SMAD1/5/8 and ID3 in the stem cell niche. Our findings suggest a novel mechanism whereby LRP2-mediated catabolism of BMP4 in the ependyma modulates the microenvironment of the SEZ and enables adult neurogenesis to proceed.
Introduction
In the adult mammalian brain, neurons are continuously generated in two germinal niches: the subependymal zone (SEZ) of the lateral ventricles (Merkle and Alvarez-Buylla, 2006; Ninkovic and Gotz, 2007) and the subgranular zone (SGZ) of the hippocampus (Ehninger and Kempermann, 2008). In recent years the morphological events underlying adult neurogenesis have been described, but the signals that orchestrate the formation of newborn neurons are incompletely understood. However, it is appreciated that the milieu of growth factors and morphogens in the neurogenic stem cell niche is critical to this process (Lledo et al., 2008).
Although there are distinct differences in the processes that regulate embryonic and adult neurogenesis in mammals, there are also remarkable similarities. Thus, fine-tuning of competing signals provided by sonic hedgehog (SHH) and bone morphogenetic proteins (BMP) are central to the regulation of neural stem cell proliferation both in the embryonic neural tube and in the adult SEZ (Balordi and Fishell, 2007; Bertrand and Dahmane, 2006; Kriegstein and Alvarez-Buylla, 2009; Lim et al., 2000; Machold et al., 2003; Sur and Rubenstein, 2005).
The low-density lipoprotein receptor-related protein 2 (LRP2) is a receptor that plays an important role in balancing signals provided by SHH and BMP4 in the developing neural tube. In the early embryo, LRP2 is expressed on the apical surface of the neuroepithelium (Assemat et al., 2005; Spoelgen et al., 2005). Absence of Lrp2 expression in mutant mice results in fusion of the forebrain hemispheres, a syndrome defined as holoprosencephaly (HPE) (Willnow et al., 1996). The underlying molecular defect has been linked to abnormal patterning of the neural tube caused by increased activity of BMP4 in the dorsal forebrain and loss of Shh expression in the rostroventral forebrain. These findings place LRP2 upstream of SHH and BMP signaling pathways in regulating early pattern formation and embryonic neurogenesis (Fuccillo et al., 2006; Spoelgen et al., 2005).
Interestingly, expression of LRP2 persists in the ventricular system of the adult brain (Zheng et al., 1994). Given its critical role in balancing key neurogenic pathways during development, we investigated its role in adult neurogenesis. We uncovered a unique expression pattern of LRP2 on ependymal cells exclusively in the lateral wall of the lateral ventricles where the stem cell niche is located. We also identified that loss of receptor expression in mutant mice results in a decline in proliferative capacity in the SEZ. Impaired proliferation of neuronal progenitor cells coincides with a distinct increase in BMP2/4 expression and enhanced activity of downstream signaling pathways, in line with a critical function for LRP2 as a negative modulator of BMP signaling, not only in embryonic but also in adult neurogenesis.
Results
LRP2 is expressed in ependymal cells of the stem cell niche
In the adult brain expression of LRP2 persists in the ventricular system (Zheng et al., 1994), but detailed information on the exact cellular location in the ventricles has been lacking. Using immunohistochemistry we detected LRP2 predominantly in the lateral ventricular system. Interestingly, the signal was restricted to the lateral and the immediate ventral-medial side of the lateral ventricle, the region of adult neurogenesis. By contrast, no LRP2 was detected on the medial side of the lateral ventricular wall (Fig. 1A). Within the lateral ventricular wall, the receptor was present exclusively on the apical surface of cells lining the ventricle but was not detectable deeper in the parenchyma in the subependymal zone (SEZ; Fig. 1A). To characterize cell types expressing the receptor we stained for S100β, an antigen highly expressed in ependymal cells. LRP2 immunoreactivity was detected close to the ependymal cells (Fig. 1B). However, in contrast to S100β, LRP2 was present exclusively in the ependyma adjacent to the stem cell niche (arrowheads in Fig. 1B), but not in S100β-positive (S100β+) ependyma at the medial side of the ventricular wall (asterisk in Fig. 1B). LRP2 was also coexpressed with sex-determining region Y-box protein 2 (SOX2) in cells of the ventricular epithelium but not the SEZ (Fig. 1C). No coexpression was observed with glial fibrillary acidic protein (GFAP; Fig. 1B). GFAP is expressed in astrocytes including B cell populations such as B1 cells, the putative stem cells of the SEZ (Mirzadeh et al., 2008).
To test whether LRP2 may also be expressed in apical extensions of B cells that reach the ventricular surface, we performed immunostaining of whole-mount preparations of the lateral ventricular surface (Fig. 2). Apical processes of SEZ B cells that reach the ventricular surface form the center of so called pinwheels, supracellular structures formed by clusters of ependymal cells (Mirzadeh et al., 2008). The pinwheel architecture of the ependyma can be visualized by staining for the cell surface marker β-catenin (Fig. 2A). In these preparations, LRP2 was detected on the apical surface of the ependymal cells but not in the extensions of B cells in the center of the pinwheel that were labeled by EGFP in a human GFAP-EGFP transgenic line (Nolte et al., 2001) (Fig. 2B).
Most ependymal cells have multiple cilia that contain CD133 (prominin 1). Although LRP2 immunoreactivity was seen on the apical surface of CD133+ cells, it did not colocalize with CD133, which suggested that the receptor is not present in the ciliary membrane (Fig. 2C; supplementary material Fig. S1). Rather, partial colocalization with γ-tubulin, which labels the basal body of the cilia, suggested that LRP2 was located in the apical membrane of the pericentriolar region (Fig. 2D; supplementary material Fig. S2).
In summary, LRP2 is expressed specifically in the lateral and ventral medial wall of the lateral ventricle, the site of adult neurogenesis. The receptor was detected on multiciliated S100β+, SOX2+ cells facing the ventricular wall, i.e. ependymal cells. By contrast, the receptor is absent from the apical processes of B cells extending into the ependymal layer.
LRP2 deficiency does not affect the structural integrity of the adult ventricular system
Mice genetically deficient for LRP2 are a useful tool to investigate the contribution of the receptor to ependymal function in the adult brain. Two different mouse models with Lrp2 gene disruption have been reported. Using homologous recombination in embryonic stem cells, we had previously generated mice carrying a receptor null allele (Lrp2−/−) (Willnow et al., 1996). In addition, a mouse line with a missense mutation at amino acid position 2721 of the LRP2 polypeptide has been obtained in an ENU screen for models with altered brain cortex formation (line 267) (Zarbalis et al., 2004). The ENU mutation results in a premature stop codon in the sequence encoding the extracellular domain of LRP2 that abolishes receptor expression (Fig. 3A). Here, we have used the Lrp2267/267 model as it has a higher rate of perinatal survival compared with the Lrp2−/− line. The difference in survival rate probably reflects distinct genetic backgrounds of Lrp2267/267 (FVB/NJ×C57BL/6J) versus Lrp2−/− (129SvEmcTer×C57BL/6N) strains. All studies were performed in receptor-deficient Lrp2267/267 animals and their gender-matched littermates. No phenotypic differences were observed between wild-type (Lrp2+/+) and heterozygous (Lrp2+/267) animals. Thus, both genotypes were used as reference and are collectively referred to as control animals (Ctr).
Because the consequence of LRP2 deficiency for structural integrity of the adult CNS has not, so far, been documented, we initially performed histo-anatomical analyses of the brains of Lrp2267/267 animals. Adult mutant mice displayed subtle features of holoprosencephaly. Although the animals exhibited some degree of facial dysmorphology such as a shortened snout (Fig. 3B), their forebrains were divided into two hemispheres (Fig. 3C). Their olfactory bulbs (OB) appeared histologically normal (Fig. 3D). In line with a mild holoprosencephalic syndrome, the lumen of the ventricular system was partially enlarged (Fig. 3E), but no discernable differences were observed in the cortical layering of the brain (supplementary material Fig. S3). Also, the architecture of the ventricular epithelium appeared normal with the typical pinwheel arrangement of ependymal cells (Fig. 4). Moreover, the ultrastructure of the ependymal cell layer in the lateral ventricles of Lrp2267/267 animals was undisturbed, as shown by immunostaining of whole-mount ventricular surface preparations for the cilia marker acetylated tubulin (Fig. 5A,B) and by scanning and transmission electron microscopy (Fig. 5C-H).
Proliferation is reduced in the SEZ but not the SGZ of LRP2-deficient mice
Defects in midline patterning during embryonic development did not grossly affect the neuroanatomy of the brain cortex or the ultrastructure of the lateral ventricles in adult LRP2 null mice. Thus, this mouse model was suitable for exploring the role of the receptor in the ventricular system of the adult brain. To test the consequence of LRP2 deficiency for adult neurogenesis, we labeled fast proliferating cells in the brains of 2- to 5-month old mice with a single intraperitoneal injection of BrdU. Twenty-four hours later, BrdU+ cells in the SEZ of control and Lrp2 mutant mice were counted (Fig. 6A,B). There were 35% fewer fast proliferating cells in the SEZ of LRP2-deficient compared with control animals (Fig. 6E). This defect in proliferation appears in adulthood since evaluation of the proliferative capacity in the lateral ventricles in mice at postnatal day 10 did not show any difference between LRP2-deficient neonates and control littermates (supplementary material Fig. S4).
Cells born in the adult SEZ migrate to the OB to form interneurons in a path known as the rostral migratory stream (RMS). No qualitative differences were seen in the appearance of this stream, when BrdU+ cells were visualized in Lrp2267/267 and control animals 4 days after BrdU application (supplementary material Fig. S5). A quantitative analysis of the total number of migratory BrdU+ cells was not performed in the RMS but in the OB. There were significantly fewer BrdU+ cells in the OB of Lrp2267/267 mice compared with controls when evaluated 2 weeks after a single BrdU dose (Fig. 6C,E). This finding was in line with the overall decrease in cell proliferation in the SEZ of these animals. The lack of BrdU+ cells in the SEZ was not caused by increased apoptosis in receptor null mice as shown by TUNEL assay (supplementary material Fig. S6).
In contrast to the lateral wall of the lateral ventricle, LRP2 is not expressed in the hippocampus. Remarkably, the proliferative capacity of progenitors in the SGZ of the dentate gyrus of receptor-deficient mice was identical to that of control animals as documented by BrdU incorporation studies (Fig. 6D,E). This observation indicated a unique role for LRP2 in neurogenesis in ependymal cells as part of the neurogenic niche of the SEZ.
LRP2 deficiency impairs cell proliferation in the SEZ
In the SEZ, B cells give rise to transit-amplifying cells (C cells) that produce type A migratory neuroblasts, the immediate progenitors of newborn neurons. The reduction in numbers of BrdU+ cells in the SEZ of LRP2 null mice may reflect impairment in expansion and/or inappropriate differentiation of precursor and progenitor cells.
To identify which cell type was affected by receptor deficiency, we labeled the various populations of SEZ precursor cells for proteins characteristic for the respective cell population. Subsequently, we quantified the number of immunopositive cells or the mean fluorescence intensity in the SEZ. We found less GFAP (Fig. 7A,E), SOX2 (Fig. 7B,E), and Lewis X (LeX; Fig. 7C,E) in the SEZ of LRP2 null mice compared with controls. GFAP labels SEZ astrocytes (B cells) whereas SOX2 marks astrocytes as well as other SEZ progenitors. LeX, a carbohydrate expressed in SEZ pluripotent neuronal stem cells, is described as one of the most specific markers for B cells contributing to the stem cell pool (Capela and Temple, 2002). Conversely, distal-less homeobox 2 (DLX2, Fig. 8A) is expressed in C cells and A cells, while doublecortin (DCX; Fig. 8B) and polysialic acid neural cell adhesion molecule (PSA-NCAM; Fig. 8C) are diagnostic of A cells. Notably, there were substantially fewer DLX2, DCX and PSA-NCAM-positive cells in Lrp2 mutants (Fig. 8D).
To further examine the reduction in various SEZ cell types, we crossed the LRP2-deficient mouse model with a reporter line expressing GFP under control of the neural precursor cell-specific nestin promoter (nestin-GFP; B and C cells). The signal for GFP in the SEZ was significantly reduced in nestin-GFP;Lrp2267/267 compared with nestin-GFP;Lrp2+/+ animals (Fig. 7D,E). In contrast to the situation in the SEZ, the number of nestin-GFP+ and PSA-NCAM+ precursor cells was similar in the SGZ of receptor-deficient mice and littermate controls (supplementary material Fig. S7).
To evaluate the impact of LRP2 deficiency on gliogenesis we analyzed the expression of the transcription factor OLIG2 in the lateral wall of the lateral ventricle. OLIG2 is specifically expressed in gliogenic progenitors in the adult SEZ (Colak et al., 2008; Hack et al., 2005). Counts of OLIG2+ cells did not reveal any significant difference between LRP2-deficient and control mice (supplementary material Fig. S8A). Undisturbed gliogenesis was also confirmed by staining for platelet-derived growth factor receptor α (supplementary material Fig. S9) that is expressed in a subset of subependymal zone astrocytes implicated in genesis of the oligodendroglial lineage (Merkle and Alvarez-Buylla, 2006).
So far, our findings indicated impaired cell proliferation in the SEZ of LRP2 null mice that equally affected the neural precursor cell population and their immediate progeny. This interpretation was confirmed by quantification of BrdU+ cells that expressed DLX2 (C cells) or DCX with or without DLX2 (A cells, neuroblasts; Fig. 9). Although the total number of BrdU+ cells was reduced in the receptor-deficient line, the proportion of individual BrdU+ C and A cell populations was unchanged in the SEZ of LRP2 null mice compared with littermate controls, suggesting that LRP2 deficiency leads to a reduction in the C cell pool and reduced proliferative activity rather than a major change in cell fate of the progenitors, e.g. towards the oligodendroglial lineage.
Loss of LRP2 increases BMP2/4 signaling in the SEZ
In the embryonic neuroepithelium, LRP2 has been proposed to act as an endocytic receptor for BMP4 (Spoelgen et al., 2005). LRP2-mediated clearance of BMP4 downregulates this signaling pathway in the forebrain and ensures neurogenesis from the ventral region of the neural tube. Given the established role of BMP4 as a negative modulator of neurogenesis in the SEZ, we wondered whether LRP2 might have a similar function in the adult brain and whether an increase in BMP signaling may underlie the neurogenesis defect in Lrp2267/267 mice. To address this question, we analyzed the BMP signaling pathway in the lateral ventricles of control and LRP2-deficient animals using immunohistochemistry. We detected a distinct increase in the expression levels of BMP2/4 in the SEZ of Lrp2267/267 compared with control mice (Fig. 10A). Enhanced activity of the BMP signaling cascade was substantiated by staining for phosphorylated forms of SMAD1/5/8 proteins, intracellular effectors of BMP signaling (Fig. 10B). Quantification of stained cell nuclei uncovered a fourfold increase in phospho-SMAD1/5/8-positive cells in the SEZ of LRP2-deficient compared with control mice (Fig. 10D). Inhibitor of DNA binding 3 (ID3) is a downstream target of BMP4 signaling (Miyazono and Miyazawa, 2002). Consistent with an aberrant increase in this morphogen pathway in LRP2-deficient mice, immunohistological signals for ID3 were also distinctly stronger in the SEZ of these animals (Fig. 10C). Quantification of ID3+ cells revealed that there were twice as many in the SEZ of Lrp2267/267 mice compared with littermate controls (Fig. 10D).
Finally to characterize the phospho-SMAD1/5/8+ and ID3+ cells in more detail, we performed co-staining experiments using antibodies against phospho-SMAD1/5/8 or ID3 with antibodies against markers for the different SEZ cell types. Strikingly, a significantly higher proportion of nestin-GFP+ and SOX2+ cells in the SEZ of LRP2-deficient mice were positive for phospho-SMAD1/5/8 or ID3 compared with controls (Fig. 11). By contrast, there were fewer DLX2+ cells (C cells) that were also positive for ID3 in both control and mutant mice and there was no statistically significant difference between genotypes (Fig. 12). These results suggest that it is predominantly the nestin+ and SOX2+ cell population (B cells) that responds to ectopic and enhanced BMP4 signaling in the SEZ.
Discussion
Our studies demonstrate an essential role for LRP2, an endocytic receptor in ependymal cells, as a negative modulator of BMP signaling in the adult stem cell niche. Loss of receptor activity results in abnormally increased activity of the BMP2/4 pathway in the SEZ. As a consequence, progenitor proliferation is suppressed and neurogenic output reduced from this niche. Neurogenesis in the SGZ, the second germinal niche in the adult mammalian brain, is not affected by receptor deficiency, supporting a distinct role of the ependyma (and LRP2) in this signaling pathway that is unique to the SEZ.
Previously, the significance of the ependyma has mainly been discussed in terms of a barrier function, protecting the brain from noxious substances that accumulate in the cerebrospinal fluid (reviewed by Del Bigio, 1995). An exciting new aspect of ependymal function was recognized when it was shown that proliferation of neural stem cells persists in the SEZ of the adult brain, a region that is intimately linked to the lateral ventricles (Lois and Alvarez-Buylla, 1993; Morshead et al., 1994; Reynolds and Weiss, 1992). How ependymal cells contribute to neurogenic processes has been a matter of debate (see Chojnacki et al., 2009). It was first suggested that ependymal cells act as neuronal precursor cells under normal physiological conditions (Johansson et al., 1999). This idea was later largely abandoned, as a role of astrocytes as the sole precursors in the SEZ was favored. However, recent reports provide new evidence that ependymal cells may act as precursor cells after brain injury (Carlen et al., 2009).
Another hypothesis proposes that rather than acting as stem cells, ependymal cells contribute to the microenvironment in the stem cell niche, which is critical for neurogenesis (Colak et al., 2008; Lim et al., 2000). This hypothesis is supported by findings that factors in the stem cell niche provide instructive signals for neurogenesis. Thus, progenitors from the SEZ grafted into other SEZs produce large numbers of neurons, but fail to do so when transplanted into non-neurogenic brain regions (Doetsch and Alvarez-Buylla, 1996; Lois and Alvarez-Buylla, 1994). Inactivation of genes expressed in the ependyma such as Numb (Kuo et al., 2006) and Sox2 (Ferri et al., 2004) impairs adult neurogenesis. However, both genes are also expressed in SEZ progenitors and no firm conclusion could be drawn about whether loss of activity in the ependyma and/or in the SEZ underlies the observed neurogenic defects. The functional elucidation of an ependymal cell surface receptor, LRP2, that is required for proper neurogenesis to occur from SEZ precursor cells now provides genetic proof for a critical contribution of the ependyma to adult neurogenesis.
It is particularly fascinating that unlike other proteins expressed throughout the ependyma (e.g. S100β, vimentin, mCD24) (Mirzadeh et al., 2008), LRP2 is not uniformly distributed throughout the ventricular system. Rather, expression is highest in the lateral ventricles where it is seen exclusively in the ependyma adjacent to the stem cell niche (Fig. 1A,B). Within the ventricular epithelium of the stem cell niche LRP2 localizes to the apical surface of multiciliated ependymal cells forming the periphery of the pinwheel structure but it is absent from the apical surface of GFAP+ B cells (Fig. 2B). Thus, LRP2 represents a unique marker that functionally discriminates ependymal cells in the neurogenic zone from ependyma elsewhere in the lateral ventricles.
Neurogenesis in the SEZ is impaired in mice genetically deficient for LRP2 (Fig. 6). Signals for nestin-GFP, SOX2, GFAP and LeX all are reduced in the SEZ of LRP2-deficient mice (Fig. 7) suggesting a negative impact of receptor deficiency on the B cell pool. Depletion of the B1 cell pool is expected to lead to a reduced number of progenitor cells. This assumption was confirmed by an overall decrease in markers for transit-amplifying progenitors and the neuroblasts (DLX2, DCX and PSA-NCAM; Fig. 8). Neurogenesis in the SGZ is not affected, demonstrating confinement of the stem cell defect to brain regions that normally express LRP2. In contrast to the reduction in transit-amplifying progenitors and neuroblasts we failed to detect a difference in oligodendrocyte precursor cell numbers in the receptor-null model (supplementary material Figs S8 and S9). Thus, LRP2 deficiency does not affect proliferation of this subset of glial progenitors.
Other than induction of apoptosis (supplementary material Fig. S6), several mechanisms might cause a decrease or re-specification of the progenitor cell population in LRP2-deficient mice. For example, impaired ability of the stem cells to self-renew (as in mice lacking the orphan receptor tailless) (Liu et al., 2008), changes in cell cycle length regulation (as in p21 mutant mice) (Doetsch et al., 2002; Kippin et al., 2005) or redirection of cell fate specification (as in mice lacking SMAD4) (Colak et al., 2008) have all been shown to affect the neural precursor pool. Whatever the exact mechanism might be, alterations in such cell intrinsic pathways in the SEZ are probably downstream of a primary defect in the instructive capacity of ependymal cells caused by LRP2 deficiency.
One of the major instructive signals in the SEZ is provided by BMPs, members of the transforming growth factor β superfamily. Both BMP4 and 7 and their receptors are widely expressed by neurons and astrocytes throughout the adult CNS (Mikawa et al., 2006; Peretto et al., 2004). Within the neurogenic niche of the SEZ phospho-SMAD1/5/8 are restricted to stem cells and transit-amplifying progenitors (Colak et al., 2008). Expression of BMPs in the neurogenic regions of the adult brain seemed counter-intuitive given their anti-proliferative potential. Indeed, overexpression of BMP7 by adenoviral-mediated gene transfer into the ependyma decreases SEZ proliferation and prevents further neurogenesis (Lim et al., 2000). Thus, raising the levels of BMP interferes with proliferation and survival of neurogenic progenitors, highlighting the importance of noggin, a potent antagonist of BMP signaling, secreted by ependymal cells. Indeed, BMP levels are crucial, because their further reduction by infusing noggin into the ventricle (Lim et al., 2000) or conditional inactivation of Smad4 in GLAST+ adult neural stem cells resulted in severely reduced neurogenesis due to the ectopic upregulation of the transcription factor OLIG2, directing the progenitors towards oligodendrogliogenesis (Colak et al., 2008).
Taken together, the findings discussed above suggest an important role for the ependyma in modulating BMP signaling, thereby providing the instructive milieu for neurogenesis in the SEZ. This hypothesis receives convincing support from our work showing that ependymal cells express LRP2, an endocytic receptor for BMPs. LRP2 has been shown to bind BMP4 (but not BMP5) in vitro (Spoelgen et al., 2005). Binding results in cellular uptake and catabolism of BMP4, conceivably reducing the amount of morphogen available to activate its cognate BMP receptors (Spoelgen et al., 2005). Inactivation of the receptor pathway in the mouse results in enhanced BMP signaling and in impairment of neurogenesis in the embryonic neural tube (Spoelgen et al., 2005). The data presented here indicate that the function of LRP2, as a negative regulator of BMP signaling, is conserved in the adult brain where lack of the receptor causes aberrant increase in BMP signaling in the SEZ (Fig. 10).
Interestingly, enhanced BMP signaling as reflected by increased phospho-SMAD1/5/8 and ID3 signals mainly affects the nestin+ and SOX2+ cell population in Lrp2267/2767 mice. When adjusted for the overall decrease in the pool of nestin+ and SOX2+ cells in the SEZ of mutant mice, a significantly larger overlap with markers of the BMP signaling pathway was observed (Fig. 11). By contrast, no specific rise in numbers of DLX2-ID3 double-positive C cell types was seen (Fig. 12). These results demonstrate that a higher share of type B cells in mutant mice than in the controls remain BMP-responsive and therefore might be impaired in their self-renewal capacity and/or in their specification and differentiation towards the neurogenic lineage.
Ultimately, enhanced BMP signals as a consequence of LRP2 deficiency coincide with a significant decrease in progenitor (C cell) proliferation, substantiating the crucial role of the ependyma in reducing BMP activity to promote SEZ neurogenesis. This model is particularly exciting as ependymal cells are unique to one of the major stem cell niches, the SEZ, and are not part of the SGZ in the dentate gyrus. Indeed, the function of LRP2 thereby highlights a key difference between the two stem cell niches that produce distinct types of neurons. Further detailed characterization of the LRP2 pathway will help in elucidating the unique processes that govern neurogenesis in these neurogenic zones.
Materials and Methods
Mouse models
An LRP2-deficient mouse strain was identified in an ENU screen for mutations that impair morphogenesis of the brain cortex. In line 267 (Lrp2267/267), a T to A transition results in a stop codon at amino acid position 2721 of the LRP2 polypeptide abolishing receptor expression (Zarbalis et al., 2004). A mouse line carrying a GFP transgene driven by the rat nestin promoter (nestin-GFP) was generously provided by M. Yamaguchi, University of Tokyo (Yamaguchi et al., 2000). A mouse model expressing EGFP under the control of the human GFAP promoter was obtained from H. Kettenmann, MDC Berlin (Nolte et al., 2001). All comparative analyses were performed in LRP2 null mice and their respective gender-matched littermate controls at 2-5 months of age.
Immunohistochemical analysis
Brain tissues were fixed in 4% paraformaldehyde in phosphate-buffered saline (PBS) overnight, followed by routine paraffin embedding and sectioning at 7-15 μm thickness (Fig. 1A, Fig. 3A,D-E, Fig. 6C, Fig 10A; supplementary material Fig. S3). Alternatively, fixed brain tissues were infiltrated with 30% sucrose in PBS for 48 hours. Thereafter, 40 μm free-floating sections were cut and stored in cryoprotectant at −20°C until further use (remaining figures unless stated otherwise). Standard immunohistochemical analysis was carried out by incubation of tissue sections with primary antibodies at the following dilutions: rabbit anti-megalin antiserum 612 (1:1000; kindly provided by J. Herz, UTSW Medical Center, Dallas, USA), mouse anti-S100β (1:1000; Sigma), mouse anti-LeX/CD15 (1:500, BD Biosciences), rabbit anti-PDGF receptor α (1:700; Santa Cruz), rabbit anti-S100β (1:2500; Swant), rat anti-prominin 1 (1:100; Ebioscience), rabbit anti-SOX2 (1:200; Chemicon), guinea pig anti-GFAP (1:1000; Advanced Immunochemicals), rabbit anti-GFP (1:400; Abcam), goat anti-DCX (1:200; Santa Cruz Biotechnology), mouse anti-PSA-NCAM (1:400; Chemicon), goat anti-BMP2/4 (1:100; Fitzgerald), rabbit anti-phospho-SMAD1/5/8 (1:500; Cell Signaling Technologies), guinea pig anti-DLX2 (1:200), rabbit anti-ID3 (1:100; Abcam) rabbit anti-Tbr1 (1:1000; Chemicon), rat anti-Ctip2 (1:200; Abcam), rabbit anti-Sox5 (1:20; Santa Cruz), or rabbit anti-FoxP2 (1:200; Abcam). Bound primary antibodies were visualized using secondary antisera conjugated with Alexa Fluor 488, 555, 546, 633, 647 (1:250; Invitrogen), FITC, TRITC, Rhodamin RedX, Cy5, Biotin-SP (1:250; Jackson Immuno Research), fluorescent conjugates of streptavidin (Invitrogen) or tyramide signal amplification kit (PerkinElmer). For double labeling experiments, the incubation with primary antibodies was performed sequentially. Apoptosis was determined on histological sections using the TUNEL assay according to manufacturer's recommendations (APO-BrdU TUNEL assay kit; Invitrogen). Immunofluorescence intensities were quantified on confocal multi-spectral laser scanning microscopes.
Electron microscopy
Samples of the lateral ventricles were dissected from adult mouse brains, fixed in 2.5% PBS-buffered glutaraldehyde, dehydrated in alcohol, osmicated, dried in a critical-point apparatus, coated with carbon, and examined in a Zeiss scanning electron microscope. For transmission EM, the tissues were fixed in 1% glutaraldehyde in PBS, postfixed in 1% OsO4 (in sodium cacodylate buffer), stained en bloc in saturated uranyl acetate, dehydrated in graded ethanol and embedded in Epon. Sections were cut with a Leica Ultramicrotome UCT, stained with uranyl acetate and lead citrate, and examined in a FEI 100 CM electron microscope.
BrdU labeling experiments
To label proliferating cells in adult or postnatal brains, BrdU (Sigma) was injected intraperitoneally at 50 mg/kg body mass once, and the animals sacrificed 24 hours (SEZ, SGZ, P10), 4 days (RMS) or 2 weeks (OB) later. For quantification of BrdU-positive cells in the adult brain, free-floating sections were treated with 2 M HCl at 40°C for 45 minutes and neutralized by washing in 0.1 M sodium borate buffer (pH 8.5) for 10 minutes. Sections were blocked in 6% donkey serum in Tris-buffered saline plus 0.3% Triton X-100 for 1 hour. Then, sections were incubated with rat anti-BrdU antibody (1:500; AbD Serotec), followed by donkey anti-rat biotin SP antibody (1:250; Jackson Immuno Research) for 2 hours, and ABC-Elite reagent (PK-6100; Vector Labs) for 1 hour. Color reaction was performed using Ni-diamino benzidine. The number of labeled cells in the various brain regions was counted on 12 histological sections of each brain (×20 objective) that covered the entire area of the lateral ventricular system and the hippocampus, from the corpus callosum in the rostral to the cerebellum in the caudal region. For brains at postnatal day 10, BrdU+ cells in the SVZ were evaluated using Cell-F microscope software (Soft Imagining System). The SVZ was defined as a region of interest (ROI) and the average grey value corresponding to immunostained nuclei recorded on five to ten SVZ coronal sections in binary color image mode. Statistical significance of data was determined using Student's t-test.
Acknowledgements
We are indebted to M. Schmeisser, P. Schrade, I. B. Kristoffersen, and K. Kampf for expert technical assistance, and to G. Lewin for critical reading of the manuscript. Studies were funded in part by the DFG, SFB 665 (to T.E.W. and A.H.).